Last released Sep 21, 2026
LLM observability and resilience decorator — semantic prompt caching, automatic retries, rate limiting, cost tracking, and a local web dashboard for any Python LLM call.